Intel, the semiconductor manufacturer, recently released its 2011 sustainability report and the news, in terms of water, wasn't good. They've failed to meet their water reduction goal: While water use per chip is down 6% from 2010, it's up 12% from 2007, the baseline year. The new goal is to achieve 2010 per chip water use by 2020; they've given themselves a higher bar and double the time to reach it.
Further, total water use in 2011 was up 11% from 2007, from 7.5 to 8.3b gallons, a trend attributed to the 'increasing complexity of manufacturing processes.' And this is despite the fact that the company has spent over $100m since 1998 on water efficiency efforts.
Intel isn't the only big company in the water intensive semiconductor industry to show disappointing water reduction results. IBM, renown for their impressive water efficiency work at their Burlington, VT plant, where advanced analytics and infrastructure upgrades reduced water use 27%, and their foray into the smart water business, reported that their total water use in 2010 was up 3% over 2009.
Samsung reports that total water use is down, but strangely omits per chip water use performance data. AMD reports that water per chip water use in 2010 was up 27% over 2009, thanks to a 'process change that required a significant increase in water used for cleaning', which would seem to indicate that manufacturing processes are confoundingly growing less efficient, in terms of water use at least.
In 2011, Texas Instruments finally got around to setting a numerical water use reduction goals, which is to reduce per chip water use by 45% by 2015. However, the company reports that total water use is up 21% due to the 'addition of new facilities' and that per chip water use is up 25%. They also make mention of the Texas drought: 'We anticipate further restrictions may be put in place in 2012 if conditions remain dry.' It would seem they have a long way to go.
Perhaps most concerning is Broadcom, which admits, in its filing to the Carbon Disclosure Project, to only collecting water use data from 5% of its facilities. That a semiconductor firm, in this day and age, fails simply to measure their water use is particularly astonishing. That old saying comes to mind: One can not manage what one can not measure.
This isn't to say that there haven't been bright spots. In its 2011 sustainability report, Toshiba reports a 29% decrease in per chip water use, well exceeding their goal of 9.5%. And of course there's IBM's efforts at its Burlington, Vermont semiconductor facility, featured in the book The Big Thirst. Water conservation technologies and advanced analytics have conspired to reduce water use by 27% and cut their energy bill by $3m, all while maintaining the same level of production. It's a study on the business case for cutting water use.
It's clear however that the industry needs to do a lot more. A perfect storm seems to be brewing. The industry is booming: the Semiconductor Industry Association reported record revenues of $299.5b for 2011 and it expects continued growth "for sensors.... in the application of MEMS, or microelectromechanical systems, which are increasingly included in smartphones, tablets, digital cameras, and numerous other consumer electronic products."
Have a look around and you'll see what they mean: everyone has an iPhone, laptop, and iPad these days. Freshwater availability is declining and demand for water is increasing. Now amidst all of this consider that one 300 millimeter water requires 2000 gallon on average.
Shareholders couldn't be blamed for being a bit nervous about this situation. Ceres, in a 2009 report, cites a JPMorgan analysis which concluded that a water-related shut down of an Intel or Texas Instruments plant could cost the companies $100-200m per quarter or $.02 to $.04 per share. In another report, Ceres concludes that the semiconductor industry is doing a poor job of communicating risk to investors and that 11 of the industry's 14 largest plants are located in the already water stress Asia Pacific region.
Common folk have reason to be fearful as well. Imagine a massive cut in semiconductor production and the impacts on the world economy: the very foundations of our information economy - phones, computers, and all of those other fantastically genius devices that make business and our daily lives hum - would be put at risk. And this is the last thing we need now, when the economy is already in the pot. Perhaps the semiconductor industry need a catastrophe, a plant shut down or some such occurrence, in order to spur real water efficiency efforts. But let's hope not.